Mesophyll conductance to CO2 in Arabidopsis thaliana


  • J. Flexas,

    1. Laboratori de Fisiologia Vegetal, Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Balears, Spain;
    Search for more papers by this author
    • *

      These authors contributed equally to this work.

  • M. F. Ortuño,

    1. Departamento Botânica e Engenharia Biológica, Instituto Superior de Agronomia, Universidade Técnica de Lisboa, Tapada da Ajuda, 1349-017 Lisboa, Portugal;
    Search for more papers by this author
    • *

      These authors contributed equally to this work.

  • M. Ribas-Carbo,

    1. Laboratori de Fisiologia Vegetal, Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Balears, Spain;
    Search for more papers by this author
    • *

      These authors contributed equally to this work.

  • A. Diaz-Espejo,

    1. Instituto de Recursos Naturales y Agrobiología, CSIC, Apartado 1052, 41080 Sevilla, Spain
    Search for more papers by this author
  • I. D. Flórez-Sarasa,

    1. Laboratori de Fisiologia Vegetal, Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Balears, Spain;
    Search for more papers by this author
  • H. Medrano

    1. Laboratori de Fisiologia Vegetal, Grup de Recerca en Biologia de les Plantes en Condicions Mediterrànies, Universitat de les Illes Balears, Carretera de Valldemossa Km 7.5, 07122 Palma de Mallorca, Balears, Spain;
    Search for more papers by this author

Author for correspondence: J. Flexas Tel: +34 971 173446 Fax: +34 971 173184 Email:


  • • The close rosette growth form, short petioles and small leaves of Arabidopsis thaliana make measurements with commercial gas exchange cuvettes difficult. This difficulty can be overcome by growing A. thaliana plants in ‘ice-cream cone-like’ soil pots.
  • • This design permitted simultaneous gas exchange and chlorophyll fluorescence measurements from which the first estimates of mesophyll conductance to CO2 (gm) in Arabidopsis were obtained and used to determine photosynthetic limitations during plant ageing from c. 30–45 d.
  • • Estimations of gm showed maximum values of 0.2 mol CO2 m−2 s−1 bar−1, lower than expected for a thin-leaved annual species. The parameterization of the response of net photosynthesis (AN) to chloroplast CO2 concentrations (Cc) yielded estimations of the maximum velocity of carboxylation (Vc,max_Cc) which were also lower than those reported for other annual species. As A. thaliana plants aged from 30 to 45 d, there was a 40% decline of AN that was entirely the result of increased diffusional limitations to CO2 transfer, with gm being the largest.
  • • The results suggest that in A. thaliana AN is limited by low gm and low capacity for carboxylation. Decreased gm is the main factor involved in early age-induced photosynthetic decline.


Gas exchange is widely accepted as a fast, nondestructive, and noninvasive method for measuring the physiological status of plants. It has recently been highlighted that such an approach needs to be further developed for Arabidopsis (Lake, 2004). This is particularly important since the use of Arabidopsis as a tool for investigating physiological, biochemical and genetic traits has grown enormously, especially after the completion of the genome sequence (Arabidopsis Genome Initiative, 2000) and because of the large variability of mutants of this species.

Currently, most gas exchange analyses are carried out using commercially available infrared gas analysis systems (Long & Bernacchi, 2003), which are equipped with cuvettes typically designed to enclose between 2 and 10 cm2 of leaf surface. The leaf area of Arabidopsis that can be inserted into one of these cuvettes is often < 2 cm2. Moreover, the close rosette growth form, short petioles and leaf fragility of Arabidopsis make this particularly difficult. As a consequence, reports of gas exchange analysis in Arabidopsis are scarce. To overcome these technical limitations, several gas exchange chambers have been specially designed to enclose whole plants (Sommerville & Ogren, 1982; Donahue et al., 1997; Lascève et al., 1997; Poulson et al., 2002; Dodd et al., 2004; Tocquin & Perilleux, 2004). Whole-plant measurements represent an optimal approach to measuring net CO2 assimilation that determines plant growth and production. However, owing to leaf superimposition and simultaneous measurements of leaves with different ages and acclimation histories, whole-plant measurements are not fully adequate to study the response of net photosynthesis (AN) to substomatal CO2 concentrations (Ci), that is ANCi curves, which provide details on leaf photosynthetic biochemistry (Farquhar et al., 1980; Long & Bernacchi, 2003). Similarly, whole-plant measurements are not adequate to use simultaneous gas exchange and chlorophyll fluorescence to estimate mesophyll conductance to CO2, that is gm (Harley et al., 1992), or to use a limitation analysis to study fundamental aspects of photosynthesis limitations in response to varying plant internal or external conditions (Grassi & Magnani, 2005). These approaches required proper leaf-level gas exchange measurements.

These technical limitations help to explain the small number of reports in the literature in which gas exchange has been measured at leaf level in Arabidopsis. Moreover, most of them have been performed with custom-built chambers specifically adapted to this plant's characteristics (Lake, 2004; Walters et al., 2004; Büssis et al., 2006). Most ANCi curves in Arabidopsis have been performed at the whole-plant level (Donahue et al., 1997; Poulson et al., 2002; Tocquin & Perilleux, 2004), which somewhat constrains their interpretation in terms of leaf biochemistry. Only recently, have Walters et al. (2004) and Büssis et al. (2006) obtained ANCi curves at the leaf level. However, mesophyll conductance to CO2 (gm) was not determined in these studies, which may have led to underestimations of the maximum velocity of carboxylation (Vc,max) and other parameters derived from the curves (Flexas et al., 2006). To the best of our knowledge, no previous estimate of gm in Arabidopsis has been published. On the other hand, pooling data from different studies suggests that net photosynthesis varies with plant age in Arabidopsis (Lake, 2004), but a limitation analysis to discern the physiological mechanisms leading to such variation has not yet been performed.

The objectives of the present work were to test a method for growing Arabidopsis plants that resulted in a leaf position suitable for measurements with a commercial gas exchange system; to use simultaneous measurements of gas exchange and chlorophyll fluorescence as well as online carbon isotope discrimination analysis to estimate mesophyll conductance to CO2 in Arabidopsis plants; and to use a quantitative limitation analysis to determine photosynthetic limitations during plant ageing.

Materials and Methods

Plant material and growing conditions

Arabidopsis thaliana L. (Heynh) ecotype Col 0 seeds were incubated in Petri plates with water at 4°C for 48 h in darkness. Day 0 was considered after stratification at 4°C. Then, two plants per pot were grown in prewetted pots filled with a substrate containing peat, perlite and vermiculite (2 : 1 : 1 v/v). The substrate was added to pots not just to fill them, but also to overpass the pot limit with an ‘ice-cream cone-like’ shape, as shown in Fig. 1(a). Plants grew with their bases on the top of the ‘cone’, and with the rosette spreading downwards over the slopes of the ‘hill’, facilitating the access of the Li-6400 (Li-Cor Inc., Lincoln, NE, USA) chamber head to basal rosette leaves, avoiding leaf area insertion and tearing problems (Fig. 1b). Pots were placed in plastic trays for subirrigation.

Figure 1.

Schematic drawing (a) and photograph (b) of an Arabidopsis plant growing in substrate overfilled pots. Rosette growing and measuring form (a) allowed easy leaf clamping with the chamber head of the Li-6400 (b), avoiding problems of leaf fragility and also fully covering the 2 cm2 area of the chamber with leaf tissue. The plant shown in the picture was 30 d old.

Plants were grown for 46 d in a growth chamber under controlled conditions (12 : 12 h photoperiod, light at 250 ± 50 µmol quanta m−2 s−1, 25°C: 20°C day : night temperature and relative humidity above 40%), watered daily and with half-strength Hoagland's solution (Epstein, 1972) applied twice weekly. Leaves from the sixth to eighth pair of the rosette were labeled. Once these leaves attained a size suitable for gas exchange measurements (i.e. on day 28 after germination), gas exchange measurements were started, and the same leaves were measured during the entire experiment. Detailed measurements were performed during two periods, corresponding to 28–31 (young plants) and 42–46 (old plants) d after germination. The rosette diameter was 10 ± 1 cm, and the number of leaves 19 ± 1, with nonsignificant differences between the two plant ages. In young plants, bolting was initiated, while in old plants the first siliques were emerged.

Gas exchange and chlorophyll fluorescence measurements

All measurements were made on the same leaves described above, at 25°C and using the 2 cm2 leaf chamber. Instantaneous gas exchange measurements at saturating light (photosynthetically active photon flux density (PPFD) of 1500 µmol m−2 s−1) and 400 µmol CO2 mol−1 air were performed several times during the experiments. Detailed gas exchange experiments were restricted to the two periods described above (i.e. 28–31 and 42–46 d after germination), and included ANCi curves and determinations of respiration in the light (Rl) and the apparent CO2 photocompensation point (Ci*) according to the method of Laisk (1977).

ANCi curves were measured, using an open gas exchange system Li-6400. Leaf gas exchange parameters were determined simultaneously with measurements of chlorophyll fluorescence using the open gas exchange system Li-6400 with an integrated fluorescence chamber head (Li-6400–40). The actual photochemical efficiency of photosystem II (φPSII) was determined by measuring steady-state fluorescence (Fs) and maximum fluorescence during a light-saturating pulse of c. 8000 µmol m−2 s−1 (Fm′) following the procedures of Genty et al. (1989):

φPSII = (Fm′–Fs)/Fm

The electron transport rate (Jflu) was then calculated as:

Jflu = φPSII× PPFD ×α

(α, a term that includes the product of leaf absorption and the partitioning of absorbed quanta between photosystems II and I). α was determined from the relationship between φPSII and φCO2 obtained by varying light intensity under nonphotorespiratory conditions in an atmosphere containing < 1% O2 (Warren & Dreyer, 2006). The resulting α was 0.454, not differing among plant ages. This value was confirmed by measuring leaf absorptance using a spectroradiometer (HR2000CG-UV-NIR, Ocean Optics Inc., Dunedin, FL, USA) as described by Schultz (1996), using the light source from the Li-6400 and making the measurements inside a dark chamber. The resulting value of 0.88 implies a light partitioning towards PSII of 0.51, well within the range of typically reported values (Laisk & Loreto, 1996).

Ten CO2-response curves were obtained from different plants for each plant age. In light-adapted leaves, photosynthesis was initiated with a CO2 concentration surrounding the leaf (Ca) of 400 µmol mol−1, and a PPFD of 1500 µmol m−2 s−1 (light saturation assessed by light response curves was near 900 µmol m−2 s−1). The amount of blue light was set to 10% PPFD to maximize stomata aperture. Leaf temperature was maintained at 25°C, and leaf-to-air vapor pressure deficit was kept between 1.2 and 1.8 kPa during all measurements. Once steady state was reached (usually 30 min after clamping the leaf), a CO2-response curve was determined. Gas exchange and chlorophyll fluorescence were first measured at 400 µmol mol−1, and then Ca was decreased stepwise down to 50 µmol mol−1. Upon completion of measurements at low Ca, this was returned to 400 µmol mol−1 to restore the original AN. Then, Ca was increased stepwise to 1800 µmol mol−1. Leakage of CO2 in and out of the leaf cuvette was determined for the same range of CO2 concentrations with a photosynthetically inactive leaf enclosed (obtained by heating the leaf until no variable chlorophyll fluorescence was observed), and used to correct measured leaf fluxes (Bernacchi et al., 2002; Long & Bernacchi, 2003). Fourteen different Ca concentrations were used for each curve in young plants. In old plants, only 10 Ca concentrations were used, to reduce the time of measurements, since these had to be performed in parallel with time-consuming online isotope discrimination measurements. Maximum velocity carboxylation (Vc,max_Ci), maximum capacity for electron transport rate (Jmax_Ci) and the velocity for triose phosphate utilization (VTPU_Ci) were calculated from these ANCi curves according to Long & Bernacchi (2003). Temperature dependence of the kinetic parameters of Rubisco were taken from Bernacchi et al. (2001), while the Γ* and Rl were obtained as already described.

Estimation of gm by gas exchange and chlorophyll fluorescence

Estimations of gm were performed using the method of Harley et al. (1992), as follows:

gm = AN/(Ci– (Γ*(Jflu+ 8(AN + Rl))/(Jflu – 4(AN + Rl))))

where AN and Ci were taken from gas exchange measurements at saturating light, and Γ* and Rl were estimated using the Laisk (1977) method. Briefly, it consisted in measuring ANCi curves at three different PPFDs (50, 200, and 500 µmol m−2 s−1) with six different CO2 concentrations ranging from 300 to 50 µmol CO2 mol−1 air at each light intensity. The intersection point of the three ANCi curves was used to determine Ci* (x-axis) and Rl (y-axis). A chloroplastic CO2 photocompensation point (Γ*) of 49 µmol CO2 mol−1 air was calculated using a simultaneous equation with mesophyll conductance, gm (Ci* +Rl/g = Γ*), according to Warren & Dreyer (2006). For this, a dummy value was given to gm in the equation which was solved by sequential iterations until convergence was found. This procedure was performed using the software package Solver of Microsoft Excel.

The values obtained for gm were used to convert ANCi curves into ANCc curves. In these, the maximum velocity carboxylation (Vc,max_Cc) and the maximum capacity for electron transport rate (Jmax_Cc) were calculated using the Cc-based temperature dependence of kinetic parameters of Rubisco by Bernacchi et al. (2002).

The obtained values of gm were compared with those obtained using totally independent methods. In young plants, because up to 14 different Ca values were used during ANCi measurements, it was possible to apply the curve-fitting method by Ethier & Livingston (2004), explained later. In old plants, we decided to test gm using the carbon isotope discrimination method, because it has been described as the most accurate method (explained later). Since this method is time-consuming (c. 2–3 h to obtain each replicate value) and, in addition, it was necessary to keep the ‘sampling time’ within a few days to allow the comparison of plant ages, the number of points in ANCi curves was reduced to 10 because of time constraints. This impaired the accurate use the curve-fitting method in these curves. Still, having two completely independent estimates of gm at each sampling time reinforces the validity of our gm estimates.

Estimation of gm by a curve-fitting method

Estimation of gm by a curve-fitting method (Ethier & Livingston, 2004; Ethier et al., 2006) requires a large number of data points to be reliable. Therefore, it was only possible to perform this estimation in young leaves. In short, the method by Ethier & Livingston (2004) fits AN-Ci curves with a nonrectangular hyperbola version of the Farquhar's biochemical model of leaf photosynthesis (Farquhar et al., 1980). This is based on the hypothesis that gm reduces the curvature of the Rubisco-limited portion of an ANCi response curve. The method has been successfully used in several studies, showing good agreement with other independent estimates of gm (Niinemets et al., 2005; Warren & Dreyer, 2006). Values of the Michaelis-Menten constant for CO2 (Kc), and oxygen (Ko) and the chloroplastic CO2 photocompensation point (Γ*) and their temperature responses used for these estimations were obtained from the Cc-based in vivo values of Bernacchi et al. (2002). The Ci cut-off point was determined based on the method proposed by Ethier et al. (2006).

Estimation of gm by carbon isotope discrimination

Instantaneous carbon isotope discrimination was measured in old leaves as previously described (Flexas et al., 2006). Leaves were placed in the 2 cm2 chamber of the Li-6400 at 400 µmol mol−1, with a PPFD of 1500 µmol m−2 s−1, and a temperature of 25°C. The flow rate was set at 150 µmol air s−1. Gas exchange parameters were measured as described with the Li-6400 system under steady-state conditions for a minimum of 45 min. Once gas exchange measurements were performed, the air exiting the cuvette was collected as follows: maintaining the leaf in the cuvette under steady state by maintaining the same conditions of light, CO2 concentration and temperature, the exhaust tube was disconnected from the match valve and connected through a series of Swagelock tube connectors to a magnesium chloride (water trap) tube and a handmade 100 ml glass flask with Teflon stopcocks (Ribas-Carbo et al., 2002). Under steady-state conditions, the air passed through the desiccant and the open collecting bottle for 15 min at a flow above 150 ml min−1, ensuring 20 full turnovers of air inside the collecting bottle before the stopcocks were closed and the bottle removed. In order to collect a reference air, the same procedure was followed with the cuvette empty.

Carbon isotope composition was determined in an isotope ratio mass spectrometer (Thermo Delta XPlus, Bremen, Germany) under dual-inlet mode. CO2 from the bottles (sample and reference) was first concentrated in a Precon loop under liquid nitrogen and then introduced in its corresponding fully expanded bellow. The bellows were then compressed to increase the signal for the m/z 44 peak to a minimum of 1000 mV to maximize the signal : noise ratio. The dual-inlet IRMS compared the isotope ratio of the sample and reference CO2 introduced in its bellows. First, the system performs a peak center on (m/z 45), then it equilibrates the sample and reference signal for the m/z 44 peak and then both isotope ratios are compared 25 times. The dual-inlet mode largely reduces measuring errors as compared with the continuous-flow mode (Ribas-Carbo et al., 2002), and hence standard deviation of the δ13C of the sample CO2 with respect to the reference CO2 was always below 0.03‰. This fact allowed the determination, with high reproducibility, of δ13C values as low as 0.3‰, and permitted obtaining precise values in Arabidopsis leaves even when using the small 2 cm2 chamber with a relatively small CO2 draw-down (see the Results section).

Carbon isotope discrimination was calculated as described by Evans et al. (1986), as follows:

Δ13Cobs = (ξ(δ13C– δ13Ce)/(1000 + δ13Co       – ξ(δ13Co – δ13Ce)))

(ξ = Ce/(CeCo); Ce and Co, CO2 concentrations of the air entering and leaving the chamber, respectively). Gas-exchange parameters AN, Ce and Co, were as determined with the Li-6400. Because of the dual-inlet comparison method used, the value of δ13Ce was equal to 0, and δ13Co was the value obtained from the isotope analysis.

Mesophyll conductance values were determined by comparing predicted discrimination with observed discrimination. Predicted discrimination (Δi) was calculated from the following equation by Evans et al. (1986):

Δi=a+ (ba)pi/pa

(a, the fractionation occurring as a result of diffusion in air (4.4‰); b, the net fractionation by Rubisco and phosphoenolpyruvate carboxylase (PEPC) 29‰; pi and pa, intercellular and ambient partial pressures of CO2, respectively).

Finally, gm was calculated from the following equation (Evans & von Caemmerer, 1996):

Δi–Δ13Cobs = (29 – 1.8)(AN/gm)/pa

where 1.8‰ is the discrimination resulting from dissolution and diffusion of CO2 in water.

Photosynthesis limitations analysis

The relative photosynthetic limitations resulting from leaf ageing were partitioned into their functional components following the method by Jones (1985) implemented by Grassi & Magnani (2005) to take into account mesophyll conductance to CO2. This approach, which requires the measurement of AN, gs, gm and Vc,max, allows the partitioning of photosynthesis limitations into components related to stomatal conductance (SL), mesophyll conductance (MCL) and leaf biochemical characteristics (BL), assuming a reference treatment in which the maximum assimilation rate, gs, gm and Vc,max can be defined. These maximum values were observed in young plants, and therefore used as the reference to calculate partial photosynthetic limitations in old plants.

Statistical analysis

Differences between means (young vs old plants) were assessed by one-way anovas (P < 0.05) performed with the SPSS 12.0 software package (SPSS, Chicago, USA).


Measuring gas exchange at the leaf level in Arabidopsis plants was performed in 10 replicates per plant age. In all measurements, the 2 cm2 area of the gas exchange cuvette was fully covered by leaf tissue, which was facilitated by the ice-cream cone structure of the growing substrate (see the Materials and Methods section, Fig. 1). At 400 µmol CO2 mol−1 air and saturating light, AN slightly increased from 14 to > 15 µmol CO2 m−2 s−1 from 28 to 34 d after germination (Fig. 2). A significant decline to values close to 13 µmol CO2 m−2 s−1 occurred between days 35 and 38. Finally, a larger decline to values of approx. 8–9 µmol CO2 m−2 s−1 was observed after day 41. A more detailed photosynthesis analysis was performed in two periods, 28–31 and 42–46 d, that is a period close to that with maximum values of AN and a period with the lowest values (Fig. 2). AN over the measuring period averaged 14.5 µmol CO2 m−2 s−1 in young plants, while stomatal conductance (gs) was 0.183 mol CO2 m−2 s−1 and Ci air was 304.2 µmol CO2 mol−1 (Table 1). AN and gs were reduced by 41 and 35%, respectively, in old plants, resulting in a nonsignificant variation of Ci (Table 1). Chloroplast CO2 concentrations (Cc) and leaf mesophyll conductance to CO2 (gm) were calculated from combined gas exchange and chlorophyll fluorescence data (Table 1). gm averaged 0.195 mol CO2 m−2 s−1 bar−1 in young plants, while Cc was substantially lower than Ci (227.2 µmol CO2 mol−1). The gm estimated using the Harley et al. (1992) method matched reasonably well the gm estimated using the curve-fitting method by Ethier & Livingston (2004), which averaged 0.225 mol CO2 m−2 s−1 bar−1 (Table 1). In old plants, gm was reduced (0.064 mol CO2 m−2 s−1 bar−1) and, as a consequence, Cc was significantly lower than in young plants. A totally independent approach based on online carbon isotope discrimination (Evans et al., 1986) was used to confirm the decline in gm (Table 1). Leaves generated a CO2 draw-down in the gas exchange chamber of 14.9 ± 1.9 µmol CO2 mol−1 air, which resulted in a change in δ13C of 0.54 ± 0.06‰. The IRMS dual-inlet system used gave a standard deviation of δ13C below 0.03‰ (see the Materials and Methods section). Therefore, it was possible to determine gm in Arabidopsis leaves accurately, even when using the small 2 cm2 chamber. AN was strongly correlated with gm when data for individual leaves were plotted (Fig. 3a). Also, gs and gm were significantly correlated (Fig. 3b).

Figure 2.

The evolution of net photosynthesis (AN) with leaf ageing (d after germination) in plants of Arabidopsis thaliana. The values shown are means ± SE of six replicates.

Table 1.  Mean values for the photosynthetic parameters analysed
 Young plantsOld plantsLimitation
  • AN, net photosynthesis; gs, stomatal conductance; gm, mesophyll conductance to CO2; Ci, substomatal CO2 concentration; Cc, chloroplast CO2 concentration; Vc,max_Ci, maximum velocity of carboxylation calculated from gas exchange on a Ci basis; Vc,max_Cc, maximum velocity of carboxylation calculated from gas exchange on a Cc basis; Jmax_Ci, maximum capacity for electron transport calculated from gas exchange on a Ci basis; Jmax_Cc, maximum capacity for electron transport calculated from gas exchange on a Cc basis; Jflu, electron transport rate estimated by chlorophyll fluorescence.

  • Values are means ± SE (n = 10). Different letters indicate statistically significant differences (P < 0.05) between young and old plants. The right column shows the results of a quantitative limitation analysis following Grassi & Magnani (2005), in which the photosynthesis limitation in old plants is referred to values in young plants. The analysis solves total limitation in AN, and partial limitations resulting from decreased gs (SL), gm (MCL) and Vc,max_Cc (BL).

  • a

    The Ethier & Livingston (2004) curve-fitting method was used only for young plants, while the Evans et al. (1986) isotopic method was used only for old plants (see the Materials and Methods section).

AN (µmol CO2 m−2 s−1)14.5 ± 0.5 a8.6 ± 0.6 bTotal: 41%
gs (mol CO2 m−2 s−1)0.183 ± 0.011 a0.119 ± 0.009 bSL: 13%
gm Harley (mol CO2 m−2 s−1 bar−1)0.195 ± 0.015 a0.064 ± 0.007 bMCL: 28%
gm Ethier/Evans (mol CO2 m−2 s−1 bar−1)a0.225 ± 0.050 a0.078 ± 0.013 b 
Ci (µmol CO2 mol−1 air)304.2 ± 3.3 a311.4 ± 3.0 a 
Cc (µmol CO2 mol−1 air)227.2 ± 6.4 a172.6 ± 9.0 b 
Vc,max_Ci (µmol m−2 s−1)58.1 ± 2.0 a34.2 ± 1.7 b 
Vc,max_Cc (µmol m−2 s−1)62.9 ± 1.5 a53.4 ± 4.0 bBL: 0%
Jmax_Ci (µmol m−2 s−1)98.8 ± 3.5 a59.4 ± 4.0 b 
J max_Cc (µmol m−2 s−1)104.4 ± 2.9 a76.9 ± 4.4 b 
Jflu (µmol m−2 s−1)110.1 ± 3.8 a80.8 ± 5.0 b 
Jmax_C: Vc,max_Ci1.70 ± 0.04 a1.74 ± 0.10 a 
Jmax_C: Vc,max_Cc1.66 ± 0.02 a1.46 ± 0.07 b 
Figure 3.

The relationship between AN and gm (a), and gm and gs (b) in young (closed circles) and old (open circles) plants of Arabidopsis thaliana. The values shown are individual replicates. In (a), the dashed line represents the regression fit between AN and gm for literature data on different plants as shown by Evans & Loreto (2000), and in (b), the dashed line represents the 1 : 1 relationship.

ANCi curves showed that both the initial slope and the maximum photosynthesis rate were higher in young plants than in old plants (Fig. 4a). Parameterization of the Farquhar et al. (1980) model of photosynthesis yielded a Vc,max_Ci of 58.1 µmol m−2 s−1 and a Jmax_Ci of 98.8 µmol m−2 s−1 in young plants (Table 1). The velocity of triose phosphate utilization (VTPU_Ci) averaged 6.7 µmol m−2 s−1 (data not shown). In old plants, Vc,max_Ci was reduced to 34.2 µmol m−2 s−1 and Jmax_Ci to 59.4 µmol m−2 s−1 (Table 1), so that the ratio Jmax_Ci/Vc,max_Ci remained similar to that in young plants (1.7). In old plants, the curves did not show limitation by triose phosphate utilization (Fig. 4a), and so VTPU_Ci could not be calculated. The maximum Ci achieved was only slightly higher in young (1586 µmol CO2 mol−1 air) than in old plants (1511 µmol CO2 mol−1 air).

Figure 4.

The relationship between AN and Ci (a), or Cc (b) in young (closed circles) and old (open circles) plants of Arabidopsis thaliana. The values shown are averages ± SE for 10 replicates per plant age.

ANCc curves differed significantly from ANCi curves (Fig. 4b). The differences in the initial slope between young and old plants were largely attenuated, but the maximum Cc achieved was more than double in young (681 µmol CO2 mol−1 air) than in old plants (307 µmol CO2 mol−1 air). Parameterization of the Farquhar et al. (1980) model of photosynthesis resulted in different values compared with the ANCi curves (Table 1). Vc,max_Cc averaged 62.9 µmol m−2 s−1 in young plants, and it was reduced only slightly (53.4 µmol m−2 s−1) in old plants. Similarly, differences between young and old plants in Jmax_Cc were lower than for ANCi curves (104.4 and 76.9 µmol m−2 s−1 in young and old plants, respectively). Clearly, Jmax_Cc matched fluorescence estimates (Jflu) much better than Jmax_Ci, particularly in old plants (Table 1). The Jmax_Cc : Vc,max_Cc ratio was similar to the Jmax_Ci : Vc,max_Ci ratio in young plants, but it decreased significantly to 1.46 in old plants. It was possible to calculate VTPU_Cc only in seven of the 10 curves for young plants, but average values did not differ from those estimated from ANCi curves (data not shown).

From values of AN, gs, gm (Harley method), and ANCc curves, a quantitative limitation analysis was performed (Grassi & Magnani, 2005). Because at near to ambient Ca (400 µmol CO2 mol−1 air) all plants were in the Rubisco-limited region of ANCc curves, Vc,max_Cc was the parameter used to estimate biochemical limitations (BL) to photosynthesis in these plants (Grassi & Magnani, 2005). The analysis revealed that AN was depressed by 41% in old compared with young plants (Table 1). Of this limitation increase, stomatal limitation (SL) accounted for 13% while mesophyll conductance limitation (MCL) accounted for the remaining 28% (Table 1). Biochemical limitations (BL) did not contribute to photosynthesis limitations in old Arabidopsis plants. Identical results were obtained when alternative estimates of gm were used (not shown).


Photosynthetic characterization of Arabidopsis leaves: diffusional and biochemical aspects

Growing Arabidopsis plants in an ‘ice-cream cone-like’ soil structure (Fig. 1) successfully permitted clamping of its leaves in the 2 cm2 area of the gas exchange and chlorophyll fluorescence chamber of the Li-6400 (Li-6400-40) covering the entire cuvette. Therefore, it was possible to perform simultaneous measurements of gas exchange and chlorophyll fluorescence in Arabidopsis leaves, as it is for other species with larger leaves using the same chamber size. Stomatal conductance (gs) values were similar to those reported for Arabidopsis in the literature (Lascève et al., 1997; Poulson et al., 2002). AN values in young plants also agreed well with the highest values reported in the literature (Walters et al., 2004), although the reported values are often much lower, ranging from 6 to 10 µmol CO2 m−2 s−1 (Lake, 2004). However, many reports do not specify the plant's age at measuring time (Lake, 2004; Walters et al., 2004; Büssis et al., 2006). Values reported for whole-plant measurements are usually much lower (Dodd et al., 2004; Lake, 2004; Tocquin & Perilleux, 2004), and this may be the result of a combination of different leaf ages and leaf superimposition leading to decreased light interception.

To the best of our knowledge, no previous report of mesophyll conductance to CO2 (gm) was available for Arabidopsis. The present results show that maximum gm was approx. 0.2 mol CO2 m−2 s−1 bar−1, and it decreased substantially with plant age. These data were confirmed using three well-accepted independent methods: the gas exchange-chlorophyll fluorescence method by Harley et al. (1992), the curve-fitting method by Ethier & Livingston (2004), and the online carbon isotope discrimination method by Evans et al. (1986). As is usually observed (Evans & Loreto, 2000), gm was well correlated with both AN and gs (Fig. 3). The relationship between AN and gm found in Arabidopsis did not extrapolate to zero, and it significantly deviated from that described by Evans & Loreto (2000) for several species. Such a deviation and nonzero intercept have already been shown in some species (Singsaas et al., 2003), and are thought to occur because the relationship is actually curvilinear, deviating from linearity particularly at gm values below 0.05 mol CO2 m−2 s−1 bar−1 (Warren & Adams, 2006). The relationship between gm and gs substantially deviated from 1 : 1, as has been found in some studies (Ethier et al., 2006) but not in others (Loreto et al., 1992). The slope of this relationship suggests that there is a shift in photosynthesis limitation from predominantly gs, when gs is higher than 0.18 mol CO2 m−2 s−1, to predominantly gm, when gs drops below 0.18 mol CO2 m−2 s−1.

While gm values found in Arabidopsis corresponded well with observed AN and gs, they were much lower than expected for a species with very thin leaves (leaf mass per area (LMA) was only 18 g m−2 in the studied plants, with no significant differences between ages). For instance, similar values of gm (c. 0.20–0.25 mol CO2 m−2 s−1 bar−1) have been described for species with thicker leaves, such as Vitis vinifera (Flexas et al., 2002; LMA = 65 g m−2), Citrus paradisi (Syvertsen et al., 1995; LMA = 108 g m−2), or even Pseudotsuga menziesii (Warren et al., 2003; LMA = 150–200 g m−2). On the contrary, much larger gm values (0.3–0.6 mol CO2 m−2 s−1 bar−1) are usually described for species having similar LMA to Arabidopsis (i.e. < 50 g m−2), such as Prunus persica (Syvertsen et al. (1995), Nicotiana tabaccum (Evans & Loreto, 2000; Flexas et al., 2006) and Phaseolus vulgaris (Singsaas et al., 2003). Therefore, we suggest that gm is limiting photosynthesis in Arabidopsis plants, at least when growing under the standard laboratory conditions used here and in most studies with this species. This could be reflecting low-light acclimation (Piel et al., 2002), although relatively low gm in thin leaves has also been observed in sun leaves of several Mediterranean herbs (Galmés et al., in press). The recent discovery that aquaporins are strongly implicated in the regulation of gm (Hanba et al., 2004; Flexas et al., 2006) provides a strong physiological background to explain the weakness of the relationship between gm and leaf structure.

The ANCi curves and derived parameters determined in young plants were very similar to those already shown for Arabidopsis at the leaf level by Walters et al. (2004) and Büssis et al. (2006), showing a maximum AN of c. 20 µmol CO2 m−2 s−1, attained at a Ci of 600–1000 µmol CO2 mol−1 air, and a photosynthesis decline associated with restricted TPU at higher Ci. When measured at the whole-plant level, however, the ANCi curves are quite different, with lower initial slopes and not showing CO2 saturation even at Ci values above 1200 µmol CO2 mol−1 air (Poulson et al., 2002; Tocquin & Perilleux, 2004). In fact, the only Vc,max_Ci values found in the literature are those by Tocquin & Perilleux (2004), determined at the whole-plant level and ranging between 21 and 45 µmol m−2 s−1, that is, substantially lower than those found in the present study. The values observed in the present study in Arabidopsis for Vc,max_Ci, Jmax_Ci and VTPU_Ci are well within the range of those typically found in the literature, and the ratios Jmax_Ci/Vc,max_Ci were very similar to those usually found in C3 plants at 25°C (Wullschleger, 1993). However, absolute rates of Vc,max_Ci, Jmax_Ci and VTPU_Ci in Arabidopsis are near the lower end of the range found in the literature, and more similar to those found in leaves adapted to low light (Piel et al., 2002), with low nutrient supply (Warren, 2004) or thick-leaved sclerophyll species (Wullschleger, 1993), than to those generally observed for herbaceous dicots (Wullschleger, 1993). These data, combined with gm, suggest that a low carboxylation capacity might also limit maximum photosynthesis in Arabidopsis.

However, it has been shown that ANCi curves underestimate Jmax and, especially, Vc,max, leading to some overestimation of the ratio Jmax : Vc,max (Flexas et al., 2002, 2006; Piel et al., 2002; Warren, 2004). The need to use estimates of chloroplastic CO2 concentration for a better photosynthetic characterization of Arabidopsis has recently been highlighted (Tocquin & Perilleux, 2004). The ANCc curves determined in the present study show that Jmax and Vc,max were only slightly, although nonsignificantly, underestimated by ANCi curves in young plants, but substantially underestimated in old plants, that is, those with lower gm. The ratio Jmax_Cc : Vc,max_Cc was actually lower than the ratio Jmax_CVc,max_Ci (Table 1). It may be argued that the estimation of these parameters are strongly affected by the choice of kinetic parameters involved in the equations (Kc, Ko and their dependence on temperature). However, since these parameters have not yet been described for Arabidopsis, any choice of parameters published for other species could be considered arbitrary. Consequently, we have selected the Ci-based and Cc-based parameters described by Bernacchi et al. (2001, 2002) for tobacco because, as both were calculated for the same species, a comparison between Ci-based and Cc-based photosynthesis parameterization would be straightforward. Hence, while there might be some uncertainty in the absolute values of Vc,max and Jmax, at least the differences between Ci-based and Cc-based values may be regarded as quite approximate. Therefore, the present results demonstrate that parameters derived from CO2-response curves of photosynthesis are underestimated when Ci is used instead of Cc, and that proper ANCc may be necessary to accurately analyse photosynthetic limitations imposed by plant ageing (see the following section) or environmental stresses.

Photosynthesis decline with plant age in Arabidopsis: diffusional and biochemical limitations

Net photosynthesis was depressed by as much as 40% from 30- to 40-d-old plants. Age-dependent decreases in photosynthesis of a similar magnitude over similar time ranges have been characterized in some herbs, such as Triticum durum (Loreto et al., 1994), Spinacia oleracea (Delfine et al., 1999) and Nicotiana sylvestris (Priault et al., 2007), but not in Arabidopsis. In perennial deciduous or evergreen plants, similar decreases are observed but over much longer time periods, such as months to years (Grassi & Magnani, 2005; Niinemets et al., 2005; Ethier et al., 2006).

It has recently been demonstrated that, in evergreen trees, most of the age-related decline in photosynthetic capacity in old leaves is the result of reduced gm and, to some extent, reductions in the activation state but not in the amount of Rubisco (Niinemets et al., 2005; Ethier et al., 2006). In deciduous perennials, by contrast, most of the age-related decline in photosynthetic capacity in autumn is the result of reduced photosynthetic biochemistry, with only a minor contribution of reduced gm (Grassi & Magnani, 2005). In annual plants, despite decreased gm in older leaves, it is usually assumed that age-related photosynthetic decline is mostly associated with the degradation of chlorophylls and the breakdown of Rubisco and other chloroplast proteins into amino acids (Friedrich & Huffaker, 1980; Ethier et al., 2006). However, Loreto et al. (1994) suggested that the age-related decline in photosynthesis in wheat may be attributed in part to a reduction in gm and in part to a decline in the amount of Rubisco. On the other hand, Delfine et al. (1999) showed in spinach that a 40% decline in photosynthesis of 50-d-old vs 22-d-old leaves was associated with a nonsignificant change in Rubisco amount and activity, a 25% decline in gs and a decline in gm of as much as 70%. However, a quantitative photosynthesis limitation analysis was lacking in these studies, which may be necessary to establish the physiological basis for age-related decline in photosynthesis in annual plants.

Here we present the first quantitative photosynthesis limitation analysis in response to leaf ageing performed in Arabidopsis and, to the best of our knowledge, in any annual species. The results clearly show that a 41% decline in photosynthesis as a result of leaf ageing was entirely caused by increased diffusional limitations to CO2 transfer, and not to biochemical restrictions. The largest limitation (28%) was the result of reduced mesophyll conductance to CO2, while decreased stomatal conductance accounted for the remaining 13%. The present results show, for the first time, that, similar to evergreen perennials, age-related photosynthesis decline in the annual plant A. thaliana is initiated by decreased mesophyll conductance to CO2. Chlorophyll and protein degradation may appear at a later stage, when photosynthesis is already largely depressed.


In conclusion, we have tested a method for growing plants of Arabidopsis that resulted in the development of a leaf position that allowed the 2 cm2 area of a commercial gas exchange and chlorophyll fluorescence system to be fully covered. It was therefore possible to perform simultaneous gas exchange and chlorophyll fluorescence in Arabidopsis leaves with the same accuracy as for other species with larger leaves. The analysis showed net photosynthesis and stomatal conductance values similar to the highest values reported for Arabidopsis. Estimations of gm using several independent methods showed maximum values of c. 0.2 mol CO2 m−2 s−1 bar−1, in agreement with the reported values of net photosynthesis but much lower than expected for thin-leaved, annual species. The values of gm were used to convert ANCi into ANCc curves, and its derived parameters, Vc,max_Cc, Jmax_Cc and VTPU_Cc, were also low, compared with those reported for other annual species. Altogether, these data suggest that low mesophyll conductance to CO2 combined with a low capacity for carboxylation limits maximum net photosynthesis in Arabidopsis thaliana.

When plants aged from 30 to 45 d, net photosynthesis declined by 40%. A quantitative limitation analysis revealed that such a decline was entirely caused by increased diffusional limitations to CO2 transfer, of which decreased mesophyll conductance was the largest. This is the first report showing that age-induced photosynthesis decline in an annual plant is initiated by decreased mesophyll conductance and not by chlorophyll degradation and protein breakdown.


This work was granted by project BFU2005-03102/BFI (Plan Nacional, Spain). M. R.-C. and A. D.-E. were beneficiaries of the Programa Ramón y Cajal (M.E.C.). M.F.O. was beneficiary of a Postdoctoral research fellowship from M.E.C. We would like to thank Dr Biel Martorell for his technical help on the IRMS and all the staff at the Serveis Científico-Tècnics of the Universitat de les Illes Balears for their help with measurements with mass spectrometer.